English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Most speaker verification tasks are studied as an open-set evaluationscenario considering the real-world condition. Thus, the generalization powerto unseen speakers is of paramount important to the performance of the speakerverification system. We propose to apply \textit {Mean Teacher}, a temporalaveraging model, to extract speaker embeddings with small intra-class varianceand large inter-class variance. The mean teacher network refers to the temporalaveraging of deep neural network parameters; it can produces more accurate andstable representations than using weights after the training finished. Bylearning the reliable intermediate representation of the mean teacher network,we expect that the proposed method can explore more discriminatory embeddingspaces and improve the generalization performance of the speaker verificationsystem. Experimental results on the VoxCeleb1 test set demonstrate that theproposed method relatively improves performance by 11.61\%, compared to abaseline system.

Learning Metrics from Mean Teacher: A Supervised Learning Method for Improving the Generalization of Speaker Verification System